A Service of Leibniz-Informationszentrum econstor Wirtschaft Leibniz Information Centre Make Your Publications Visible. zbw for Economics Schmeling, Maik Working Paper Institutional and Individual Sentiment: Smart Money and Noise Trader Risk Diskussionsbeitrag, No. 337 Provided in Cooperation with: School of Economics and Management, University of Hannover Suggested Citation: Schmeling, Maik (2006) : Institutional and Individual Sentiment: Smart Money and Noise Trader Risk, Diskussionsbeitrag, No. 337, Universität Hannover, Wirtschaftswissenschaftliche Fakultät, Hannover This Version is available at: http://hdl.handle.net/10419/22449 Standard-Nutzungsbedingungen: Terms of use: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated licence. www.econstor.eu Institutional and Individual Sentiment: Smart Money and Noise Trader Risk Maik Schmeling† Discussion Paper No. 337 May 2006 ISSN 0949-9962 Abstract: Using a new data set on investor sentiment we show that institutional and individu- al sentiment proxy for smart money and noise trader risk, respectively. First, using bias-adjusted long-horizon regressions, we document that institutional sentiment fo- recasts stock market returns at intermediate horizons correctly, whereas individuals consistently get the direction wrong. Second, VEC models show that institutional sentiment forecasts mean-reversion whereas individuals forecast trend continuati- on. Finally, institutional investors take into account expected individual sentiment when forming their expectations in a way that higher (lower) expected sentiment of individuals lowers (increases) institutional return forecasts. Individuals neglect the information contained in institutional sentiment. JEL-Classification: G11, G12, G14 Keywords: investor sentiment, predictive regressions, noise trader, smart money We would like to thank two anonymous referees and Lukas Menkhoff for very useful comments and especially Manfred Hubner¨ for kindly providing the data. †Maik Schmeling, Department of Economics, University of Hannover, K¨onigsworther Platz 1, D-30167 Hannover, Germany, [email protected] 1 Introduction This paper empirically investigates two questions that have been subject to a large amount of research effort and debate in financial economics, namely (i) does investor sentiment matter for stock returns, and (ii) what is the difference between individual and institutional investors in financial markets? While it seems to be generally accepted that institutions differ from indviduals due to their size and sophistication (Kaniel, Saar and Titman, 2005) there is considerable disagreement in which way these two investor groups differ and how this difference affects market processes like price formation and liquidity provision. Several studies find institutions to be informed investors or ”smart money”1 (e.g. Chakravarty, 2001, Jones and Lipson, 2004, Sias, Starks and Titman, 2004) and individual investors to be irrational noise traders or ”dumb money” (Frazzini and Lamont, 2005, Bange, 2000). However, this evidence is accompanied by the finding that institutions deli- berately herd in and out of stocks (see e.g. Nofsinger and Sias, 1999, Sias, 2004) and that they heavily rely on momentum-style strategies (Badrinath and Wahal, 2002, Griffin, Harris and Topaloglu, 2003, Temin and Voth, 2004). Furthermore, ”dumb” individuals seem to earn excess returns by providing immediacy for institutional trading demands at high frequencies (Kaniel, Saar and Titman, 2005, Campbell, Ramadorai and Voulteenaho, 2005). Therefore, the evidence from real-world tra- ding data so far is not conclusive regarding the role of these two investor groups. The question whether sentiment, or the mood and expectations of investors, matter for stock returns is more controversial and supporters from the behavioral side (e.g. Shiller, 2003) and critics from the rational camp (e.g. Fama, 1998) have arguments in favour of this view or against it. While theoretical models have early incorporated 1We refer to institutions as smart money in the sense of informed investors (e.g. Campbell and Kyle, 1983) and not in the narrower context of mutual fund flows only as in Zheng (1999). 2 the existence of noise traders into equilibrium asset pricing (Kyle, 1985, DeLong et al., 1990), empirical evidence on the relevance of investor sentiment does not provide clear findings (see e.g. the polar results in Brown and Cliff, 2005 and Wang, Keswani and Taylor, 2006). We tie up these two questions and investigate whether sentiment of institutionals and individuals matters for aggregate stock market movements and whether the influence of sentiment of these two groups is systematically different. Using a new data set that covers both institutional and individual investors we find, first, that sentiment matters for several stock markets around the world and over intermedia- te horizons of up to one year. Second, there is a sharp difference between the two investor groups. Institutional investor sentiment forecasts stock returns correctly on average. Individual sentiment negatively predicts market movements in a way that is consistent with the hypothesis that overoptimistic (-pessimistic) noise traders drive markets away from intrinsic values. Third, in line with these findings, institutio- nal investors become more pessimistic (optimistic) when they expect individuals to become more optimistic (pessimistic). Therefore, our contribution to the literature is twofold. We first employ a new data set that covers genuine investor sentiment from a weekly survey, twice-separated on individual and institutional investors as well as on short and medium forecasting horizons based upon several major stock markets around the world. This data set allows us to analyze investor sentiment while controlling for factors like the geo- graphical distance of a market relative to the home country, forecast horizon and sophistication of investors. This is new to the literature since previous studies ha- ve to rely on proxies for (mostly institutional) sentiment (e.g. Neal and Wheatley, 1998, Bodurtha, Kim and Lee, 1995, or Wang, Keswani and Taylor, 2006) or rather examine sentiment of individual investors for the US market exclusively (Kumar and Lee, 2004 or Lee, Jiang and Indro, 2002). 3 Second, we contribute to the literature by directly extending a new empirical mo- delling approach from Brown and Cliff (2005) to the case of two investor groups. Earlier studies employing sentiment data have focussed on short run predictability in first or second moments (Lee, Jiang and Indro, 2002, Wang, Keswani and Taylor, 2006). Following Brown and Cliff (2005) we investigate longer term effects of sen- timent on stock markets since the building up of excessive optimism or pessimism, i.e. sentiment, is likely to be a persistent process whose effects on stock prices would be hard to detect over short horizons. Whereas Brown and Cliff limit their analysis to individuals, we jointly analyse the impact of both individuals and institutions on stock prices and complement their approach with further analyses that all point to the main result of this paper: individual sentiment is a proxy for noise trader risk and institutions seem to be smart money. The rest of the paper unfolds as follows: the next section derives testable hypotheses from earlier studies and section 3 describes the data set. Section 4, 5 and 6 show empirical results and section 7 concludes. 2 Hypotheses and earlier literature One of the basic issues related to studies of investor sentiment deals with the question whether sentiment contains unique information for asset pricing. Indeed, there is lots of evidence that investor sentiment, moods or the awareness of investors for certain topics affect conditional moments of equity returns. This includes among others the high-volume premium documented by Gervais and Kaniel (2001), index additions and deletions (see e.g. Chen, Noronha and Singal, 2004) or rumors and talks in internet chatboards investigated by Antweiler and Frank (2003). On a theoretical level, the model of Barberis, Shleifer and Vishny (1998) gives room 4 for systematic under- and overreaction of stock returns due to shifts in investor sen- timent. Using several proxies for investor sentiment Neal and Wheatley (1998) and Baker and Wurgler (2006) find that sentiment proxies heavily affect the cross-section of stock returns, e.g. that they affect the size effect or the relative prospects of diffe- rent groups of stocks sorted by characteristics like volatility and dividend payments. These results seem to carry forward to the realm